Astrometric detection of extrasolar planets: Results of a feasibility study with the Palomar 5 meter telescope

The detection of extrasolar planets around stars like the Sun remains an important goal of astronomy. We present results from Palomar 5 m observations of the open cluster NGC 2420 in which we measure some of the sources of noise that will be present in an astrometric search for extrasolar planets. This is the first time that such a large aperture has been used for high-precision astrometry. We find that the atmospheric noise is 150 mu as hr(1/2) (1 mu as = 1 micro-arcsecond) across a 90 '' field of view and that differential chromatic refraction (DCR) can be calibrated to 128 mu as for observations within 1 hr of the meridian and 45 degrees of zenith. These results confirm that a model for astrometric measurements can be extrapolated to large apertures. We demonstrate, based upon these results. that a large telescope achieves the sensitivity required to perform a statistically significant search for extra solar planets. We describe an astrometric technique to detect planets, the astrometric signals expected, the role of reference stars, and the sources of measurement noise: photometric noise, atmospheric motion between stars, sky background, intrumental noise, and DCR. For the latter, we discuss a method to reduce the noise further to 66 mu as for observations within 1 hr of the meridian and 45 degrees of zenith. We discuss optimal lists of target stars taken from the latest Gliese & Jahreiss catalog of nearby stars with the largest potential astrometric signals, declination limits for both telescope accessibility and reduced DCR, and galactic latitude limits for a sufficiant number of reference stars. Two samples are described from which one can perform statistically significant searches for gas giant planets around nearby stars. One sample contains 100 ''solar class'' stars with an average stellar mass of 0.82 M(.); the other maximizes the number of stars, 574, by searching mainly low-mass M stars. We perform Monte Carlo simulations of the statistical significance of the expected results by using measured and estimated noise quantities. We show the semimajor axis parameter spaces that are searched for each star and how an increase in the length of the observing program expands these spaces. The search over semimajor axis parameter space relates to the theory of gas giant planet formation.